Recent studies have identified a subset of natural killer cells that express T cell receptors (NKT cells). These cells have been proposed to serve a sentry function in the liver and are likely to be involved in early detection and elimination of tumor cells metastasizing to the liver. However, the nature of this sentry function is poorly understood. A major obstacle to elucidation of the sentry function of NKT cells in the liver is the lack of appropriate methodology for quantifying the migratory behavior of the NKT cells in the hepatic sinusoids. Previous studies have proposed that the movement of NKT cells through the sinusoids is essentially random in the absence of a target for detection. We hypothesize that the presence of tumor cells in the liver causes the NKT cell motion patterns to become progressively less random and more directed as a function of proximity to the tumor cell. In order to test this hypothesis, we propose three aims: 1) Develop a methodology that will allow automatic detection and tracking of liver NKT cells and tumor within the hepatic lobule in vivo. 2) Develop a methodology that will allow automatic characterization of the NKT cell's motion behavior with respect to tumor including extravasation. 3) Test the hypotheses of the relationship between the NKT cell's motion behavior and tumor based on a large amount of motion behavior data of NKT cell. In order to accomplish these aims, we have established a collaboration between an expert in in vivo imaging of the liver and an expert in image analysis and computational methods. Successful completion of this project would provide a novel approach to the analysis of sentry behavior of these immune cells in vivo as well as elucidate the mechanisms by which NKT cells locate tumor cells in the liver in vivo. [unreadable] [unreadable] [unreadable]